[Bf-committers] sequencer imbuf usage

[Campbell Barton]

Yesterday William noticed a bug in the sequencer that is more obvious
with flipping but happens with all image operations - (filtery, multiply
etc)

When you import 1 image and then duplicate it, they share an ImBuf even
if the settings are not the same, so the settings from the first
instance are used only.

Had a quick look at fixing this but theres no obvious solution, maybe
store settings in the imbuf, and do a comparison whenever checking if
the existing imbuf is usable, 
or we could store a pointer to the last used image strip in the userdata
pointer (would need to check pointers whenever deleting of course)
...or just have every image strip use its own imbuf's could save all the
hassle with imbuf comparison and the expense of some memory.

This isnt slowing productivity or anything, would just be nice to have
fixed.

On Mon, 2007-10-15 at 18:39 +1000, Campbell Barton wrote:
> Sorry about this bug, I realized the header was bad after adding this
> but had no svn access over the weekend, the artists have some more
> requests but Ill pass them by you first.
> 
> - Cam
> 
> On Fri, 2007-10-12 at 08:37 +0200, Peter Schlaile wrote:
> > Revision: 12254
> >           http://projects.blender.org/plugins/scmsvn/viewcvs.php?view=rev&root=bf-blender&revision=12254
> > Author:   schlaile
> > Date:     2007-10-12 08:37:20 +0200 (Fri, 12 Oct 2007)
> > 
> > Log Message:
> > -----------
> > == IMBuf fix ==
> > 
> > Fixed Campbell's patch for IMB_flipx:
> > - header file declaration is corrected
> > - reordered loops to make things faster (less cache misses 
> >   and no tests for float-buffers on a _per pixel basis_!)
> > 
> > (Campbell: it'd be nice, if you could check with me before patching
> > the sequencer. Or at least pay attention to compiler warnings :) )
> 
> 
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers